Effects of Smoking on Body Weight among University Students

Aim 

The aim of this study is to examine the effect of smoking on Body Mass Index among university students. 

Hypothesis 

Several mechanisms are recommended to describe the relationship between smoking and Body Mass Index (BMI). Past studies support the idea that people who smoke weigh less and that their BMI is lower compared to that of non-smokers. It is suggested that smoking caused a decrease in weigh through causing an increase in energy expenditure and reduced appetite. Generally, smokers have a low BMI that those who have never smoked. This relationship is systematic in large cohorts and Mendelian randomization studies that test molecular methods and causality behind body mass index and smoking. Causal effects of smoking are also supported by the proof that smokers who quit smoking start gaining weight. Therefore, the current study’s hypothesis is that students who smoke will have a body mass index (BMI) that is lower than that of non-smokers. 

Background 

Although the negative effects of smoking are not questionable, the high likelihood of gaining weight after quitting smoking has raised concerns related to the impact of policies against smoking on obesity (Jacobs 1). The relationship between smoking and body weight is now a critical issue in obesity studies. However, the growing evidence provides some conflicting findings. However, most studies purport that individuals who smoke cigarettes weigh less than those who do not smoke. In addition, some claim that former smokers do not weigh more than the population that has never smoked. Other studies argue that a direct relationship exists between smoking and a significant increase in weight. Others claim that a significant reduction in cigarette smoking causes a very small increase in obesity rates. Concerns on the relationship between smoking and body weight have increased because smoking and obesity are some of the primary risk factors for most non-communicable illnesses (Piirtola et al. 2). Both conditions contribute to a higher possibility for premature death and increased expenditure in healthcare. Smoking cigarettes is the second leading risk factor for increased mortality throughout the world. 

Smoking cessation is linked with a reduced risk of cardiovascular disease and cancer. An increase in weight after quitting smoking is, in most cases, cited as a reason to not stop cigarette smoking, particularly among female smokers. Although smoking rates have decreased significantly across the globe in the past few decades, body mass index has substantially increased throughout the same period. A common perception by many people is that quitting smoking causes an increase in body mass index. While smoking is associated with a decreased body mass index for adults, most studies suggest that the trend is not seen among young smokers aged between sixteen and twenty-four years. The weight control impact of cigarette smoking is said to be lacking in consistency among people in their developmental period or in the first stages of use (Jacobs 1). Some studies suggest that smoking has an anti-estrogenic impact on young people, which could reduce fat hence cause weight loss. However, effects across genders seem to vary, with females being most likely to begin cigarette smoking and maintain the effects related to weight. The objective of this study, therefore, is to examine the effects of smoking on young people. 

Significance 

The findings of this study will be of importance in the development of anti-smoking policies in the university, students, the general population, and relevant organizations. Most students at UCR are young people, with some of them being smokers. The study will shed light on the impact of smoking on body mass index among young people hence lead to an informed decision on smoking cessation, interventions, and anti-smoking policies. 

Research Design and Methods 

Design 

Primary data will be collected for the study. A cross-sectional research design will be applied. A cross-sectional research design will be appropriate for the study because the researcher intends to collect information from the participants without manipulating the research environment. In addition, the design allows a comparison of data from different groups and variables at the same time (Levin 24). Also, a cross-sectional method id appropriate for the study because of time limits. Cross-sectional research is usually conducted at a one-time point or in a very short period (25). They are used to approximate the prevalence of a particular outcome of interest in the study population. They make it possible to collect data on different characteristics. As a result, cross-sectional studies give a snapshot of findings and the related characteristics during a certain period (Levin, 24). A cross-sectional study is applied when the aim of the research is descriptive, in most cases, as surveys. As a result, surveys will be used to collect primary data from study participants. Survey questions will be sent to the participants via e-mail. A time frame of one month will be given for all the participants to have the surveys completed and submitted to the researcher. Participants will receive an e-mail reminder to ensure a high response rate. It is expected that 98 percent of all the respondents will complete the survey questions. The investigator alone will administer the surveys. Since they are delivered via e-mail, there will be no need to recruit other people to administer the surveys. 

The advantages of a cross-sectional research design are that it is not expensive. It takes a short period to implement and can approximate the prevalence of outcomes because the sample used is derived from a larger population (Levin 25). Also, most results and different risk factors can be evaluated, useful in hypothesis generation, and allows follow-up. However, the research design has several limitations that include difficulties in making causal inferences, and the fact that it only provides a snapshot and circumstances may give varying results if a different framework is incorporated. Furthermore, cross-sectional designs have a high risk of bias (Levin 25). A cross-sectional design was found appropriate for the study because it will allow the collection of primary data on a new sample that represents the whole population over a short period. A longitudinal study design would not be appropriate and suitable for the study because it takes a long period to complete as variables are repeatedly observed over a long period. Therefore, real-time results may not be generated. The advantages of using mail surveys are that they are easy, cost-efficient, and since there is no interviewer, participants may share more. The disadvantages are that response rates are, in most cases, lower than expected; they are not appropriate for participants whose literacy levels are low, and participants cannot be asked more questions for clarity because there is no interviewers. 

Sample 

The population to be studied is university students. The study population will be sampled through stratified random sampling where they will be grouped into four strata according to a year of study (first, second, third, and fourth years). In stratified random sampling, the stud population is grouped into units from which a random sample is selected (Nguyen et al. 1). The groups formed are known as strata, and from each stratum, random sampling is applied to identify a per stratum random sample. All the samples derived from each stratum are then combined to form a stratified random sample. Therefore, random samples will be identified from each stratum of university students (1 st , 2 nd , 3 rd , and 4 th -year students). The standard deviation for each stratum will be calculated to determine the allocation for each. The four random samples will then be combined to make one stratified random sample that will participate in the study. Stratified random sampling allows emphasis on particular strata over the rest by controlling the assignment of sample size (Nguyen et al. 1). 

A stratum that has a higher standard deviation, for example, can be allocated a larger size over the others that have a small standard deviation. The key strength of stratified random sampling is that the sample can be used in answering questions for aggregates of the whole population and those of population subsets, which are defined by choosing predicates offered when asking questions (Nguyen et al. 1). Stratifying target populations is commonly used in survey sampling. A major limitation with stratified random sampling is related to the use of large amounts of incoming data. The present stratified random sample strategies are primarily offline, and they assume that the required data is present before sampling begins. As a consequence, a study that uses a stratified random sample may encounter major challenges trying to adapt to new data (Nguyen et al. 1). New data may require re-computation of the samples used all over again. 

Only university students at UCR will be included in the study. A sample of 400 participants will be used for the study. A list of the student population in the university based on years of study will be obtained from UCR administration for sampling. Invitation letters will be sent via e-mail to participants. Those who accept the invitation and confirm participation will be included in the study. To encourage participation, participants will receive two US dollars as an incentive. Another advantage of stratified random sampling is that the possibility of human bias is significantly reduced. Therefore, a stratified random sample is highly representative. Since probability methods are used in stratified random sampling, the statistical conclusions drawn from the data gathered is considered valid. 

When compared to other random sampling methods, stratified random sampling is superior in that it enhances even distribution of units across the population being studied. In addition, if the samples obtained from different strata are equal in size, more precision can be obtained; hence a small sample size can be used in such a study. The disadvantages are that stratified random sampling can only be conducted if an exhaustive list of the study population can be accessed (Nguyen et al. 2). It might be challenging to access a complete list of the study population because of privacy policies, among other reasons. 

Measures 

The X variable, which is the causal or predictor variable of the study, is smoking status. The X variable survey question(s) and response options are : 

Question: Have you ever smoked? 

Response: Yes 

Question: For how long have you been smoking? 

Response: For 6 months now 

Question: How many times do you smoke in a day? 

Response: I smoke at least twice every day. 

Meaning of higher values of X: A higher value of Smoking Status refers to the person who smokes 2 or more times in a day. 

Source of X : V ariable created by the investigator 

The Y variable, which is the outcome variable, will be body mass index (BMI). 

Y survey question(s) and response options are : 

Question: Do you know your BMI level? If yes, could you indicate? 

Response: Yes, my BMI is 19. 

Meaning of higher values of Y: A higher value of BMI refers to the larger value of BMI. 

Source of Y: Variable was created by the investigator. 

The indicators will be body mass index for body weight, smoking status an indicator for smoking or non-smoking, and a number of times smokers in the study smoke cigarettes. 

Analysis 

The study will test the hypothesis “ S tudents who smoke will have a lower body mass index (B MI) than those who do not smoke.” Data analysis will be conducted using SPSS. BMI averages of smokers and non-smokers will be calculated. A descriptive study of data will be conducted to establish the different percentiles in both smokers and those who do not smoke. Homogeneity in the data will be assessed to determine the tests to be used. It is hypothesized that the student T-test will be appropriate in comparing the BMI of smokers and non-smoking students. Chi-square test will be used to explore the variations in underweight, normal weight, overweight, and obesity. The odds ratio will be adjusted using multiple logistic regression. It is hypothesized that the results of the analysis will indicate that the average BMI of students who smoke is lower compared to the BMI of students who do not smoke. 

The student t-test is applied when comparing two independent groups (Breakwell 487). The test determines the significance of the differences that exist between the means of two groups of data. Therefore, the differences between the means relative to identified random variations in the two sets of data are compared (Breakwell 487). Two types of t-tests can be applied depending on if the data sets are independent or paired. Data that is independent is derived from two separate groups, while paired data is gathered from the same group at various intervals/times or in different environments. In analyzing the data on the effects of smoking on body mass index, a t-test for independent data will be used. This is because the data will be collected from two separate groups, that is, smokers and non-smokers. 

A Chi-Square test is used to test the relationship that exists between categorical variables in a study (Breakwell, 493). For a chi-square test, the null hypothesis is always that a relationship does not exist between categorical variables. In the study on the effects of smoking and lower BMI, the question to be answered will be. ‘Is there a significant relationship between smoking and lower body mass index among university students?’ The test is, in most cases, used to assess independence tests in cross-tabulation where two categorical variables are distributed simultaneously. Tests of independence determine if a relationship exists between variables by comparing the seen series of responses to the expected pattern if variables are, in reality, independent. The calculation of the chi-square statistic will enable comparison against a crucial value in the distribution. The test will establish if there are significant differences between cell counts (Breakwell 494). 

Ethical Issues 

The ethical issues that may arise in the study are related to privacy. Obtaining the list of students in the university and their e-mail addresses involves privacy concerns. The human subjects involved will be protected by ensuring that the information gathered from them is not exposed. The data will be anonymous as well as confidential. The surveys will not require them to include their names and other sensitive personal information for privacy issues. Debriefing will not be required in the study. Also, deception will not be involved while conducting the study. There are no risks to participants identified in the study. The benefits of participation to participants are the incentives as well as insights into study findings, which could be of help, especially to smokers. Personal identifying information will not be used in reporting the findings since the data will be aggregated. The above choices related to ethical issues will ensure the confidentiality and privacy of the participants. However, the fact that the administration can access the data is an ethical limitation for the study. 

Works Cited 

Breakwell, Glynis M., et al.  Research Methods in Psychology . SAGE, 2006. 

Jacobs, Molly. “Adolescent smoking: The relationship between cigarette consumption and BMI.”  Addictive Behaviors Reports , vol. 9, 2019, p. 100153. 

Levin, Kate A. “Study design III: Cross-sectional studies.”  Evidence-Based Dentistry , vol. 7, no. 1, 2006, pp. 24-25. 

Nguyen, T. et.al. “Stratified Random Sampling from Streaming and Stored Data.”  Welcome to ISU ECpE • Electrical and Computer Engineering • Iowa State University , www.ece.iastate.edu/snt/files/2019/01/sss-edbt-2019.pdf. 

Piirtola, M., et.al. “Association of Current and Former Smoking With Body Mass Index: A Study of Smoking Discordant Twin Pairs From 21 Twin Cohorts.”  PubMed , www.ncbi.nlm.nih.gov/pubmed/30001359. 

Appendix 

Implementation Timeframe 

Date	Activity
5 th to 20 th August, 2020	Preparation and IRB
3 rd t0 31 st September, 2020	Participants recruitment
2 nd to 30 th October, 2020	Data collection
5 th to 26 th November, 2020	Data analysis
12 th December, 2020	Reporting